Yesterday acting Centers for Disease Control and Prevention (CDC) Director Jim O’Neill signed a decision memo adopting the most significant weakening of childhood vaccine recommendations in modern American history. Vaccines against hepatitis A, hepatitis B (for infants born to mothers who test negative), rotavirus, meningococcal disease, and influenza were moved from routine recommendation to “shared clinical decision-making.” 

The category sounds reasonable, even collaborative. It is neither. And it will put children’s health and lives at risk.

This decision did not emerge from the Advisory Committee on Immunization Practices (ACIP), which typically guides the CDC’s vaccine decisions. It did not follow the standard process of public deliberation, evidence review, and expert input. It came directly from Health and Human Services (HHS) Secretary Robert F. Kennedy Jr., justified by a 33-page assessment authored by Tracy Beth Høeg, MD, PhD, acting director of the Food and Drug Administration’s Center for Drug Evaluation and Research, and Martin Kulldorff, PhD, chief science and data officer for the Assistant Secretary for Planning and Evaluation. Both are longtime critics of the existing vaccine schedule who now hold the positions to dismantle it.

The category of shared clinical decision-making has a legitimate history. Understanding how it has been repurposed requires understanding what it was designed for and what it is being turned into.

The original intent

Shared clinical decision-making emerged as a recommendation category around 2015, initially for meningococcal B vaccines in adolescents, for situations of genuine clinical equipoise. These are cases in which individual factors meaningfully shift the risk-benefit assessment and population-level benefit is uncertain. 

The clearest example is human papillomavirus (HPV) vaccination for adults aged 27 to 45, which ACIP placed in this category in 2019. By that age, most people have already been exposed to HPV, and vaccine effectiveness is lower. But some individuals, say, someone entering new sexual partnerships after a divorce who was never vaccinated, might still derive significant benefit. For that population, a blanket recommendation makes little sense, but foreclosing access entirely would harm the subset who could benefit. Shared decision-making threads the needle: The vaccine remains available and covered, but the default shifts from “vaccinate” to “discuss.”

This is a coherent framework. The problem is that it is now being applied where it does not belong.

The scope of the change

Yesterday’s decision reorganizes the childhood vaccine schedule into three tiers. The first tier, vaccines recommended for all children, retains measles, mumps, rubella, diphtheria, tetanus, pertussis, polio, Hib (Haemophilus influenzae type B), pneumococcal disease, HPV (reduced to one dose), and varicella (chickenpox). 

These are what the Høeg-Kulldorff assessment calls “consensus vaccines.” Most were defined by their use in the majority of 20 peer nations surveyed, though varicella was retained despite lacking majority support, ostensibly to avoid the epidemiologic risks of halting a mature vaccination program.

The administration can claim it has not eliminated recommendations while functionally undermining them.

The second tier, vaccines recommended for high-risk groups, includes RSV prevention, hepatitis A, hepatitis B (for infants of positive mothers), meningococcal vaccines, and dengue.

The third tier, shared clinical decision-making, now encompasses rotavirus, influenza, and COVID-19 (which had been moved to this category in May 2025), hepatitis A, hepatitis B (for infants of negative mothers), and meningococcal vaccines.

The framing is strategic. No vaccine has been removed from the schedule entirely. All remain covered by insurance. The administration can claim it has not eliminated recommendations while functionally undermining them. 

The Høeg-Kulldorff assessment emphasizes that, unlike in Europe, where non-recommended vaccines are often not covered, moving these vaccines to shared decision-making in the United States ensures they “continue to be covered without cost sharing.” This is being deployed as a shield against accusations of removing access.

Why framing matters

Some will argue this is a distinction without a difference, that vaccines remain available and covered regardless of recommendation category, and that patients who want them can still receive them. This misunderstands how recommendations function in practice.

Routine recommendations establish defaults. They trigger automatic prompts in electronic health records. They shape what providers discuss and when. They send a signal about what the weight of evidence supports. Critically, they enable standing orders, the protocols that allow nurses and pharmacists to vaccinate without a specific physician examination or direct order for each patient.

Shared decision-making voids all of this. By mandating an individual deliberation for every dose, the designation creates a logistical bottleneck. A physician must be physically involved in every vaccination decision, drastically reducing throughput in busy clinics.

Beyond logistics, the designation forces providers to abandon the “presumptive approach,” the evidence-based communication strategy whereby a provider simply announces that shots are due. Research shows this is the single strongest driver of vaccine acceptance. Shared decision-making mandates an open-ended negotiation that signals ambivalence to parents.

More than 100 million Americans lack usual access to primary care. For them, the instruction to “discuss with your clinician” is not neutral guidance. It is a barrier. The mother who delivers at 2:00 a.m. without having established prenatal care, or whose positive hepatitis B test result was lost in the transfer between clinic and hospital, does not have a long-term relationship with a physician to navigate a shared decision-making conversation. The birth dose exists specifically as a safety net for her infant.

The research on this point is consistent: Shared decision-making recommendations correlate with lower uptake than routine recommendations. When you move a vaccine out of the routine category, you strip away the infrastructure that makes vaccination possible in a busy, fragmented health care system.

The Denmark illusion

The intellectual scaffolding for this decision rests on a single premise: that the United States should align its vaccine schedule with “peer nations,” particularly Denmark. The Høeg-Kulldorff assessment invokes Denmark repeatedly as a model of restraint and trust-building.

The comparison does not hold.

Denmark has 6 million people, a homogeneous population, universal healthcare with guaranteed rapid access, and a robust public health infrastructure that reliably screens every pregnant woman for hepatitis B and ensures follow-up. The country uses a Civil Registration System that digitally links maternal test results to the infant’s health record, ensuring virtually zero loss to follow-up. 

The United States has 330 million people, profound health disparities, 27 million uninsured, and a fragmented system where approximately 15% of pregnancies go unscreened for hepatitis B. Even among women who test positive, only 35% receive recommended follow-up care. The US schedule was designed as an “opportunistic” safety net precisely because the infrastructure for Danish-style precision does not exist here.

The research on this point is consistent: Shared decision-making recommendations correlate with lower uptake than routine recommendations.

The United States tried targeted hepatitis B screening in the 1980s. It failed because 35% to 65% of infected mothers had no identifiable risk factors. We switched to universal infant vaccination in 1991 because selective screening could not reliably identify every infant at risk.

Denmark is also an outlier among wealthy nations, not a model. The United Kingdom recommends 15 vaccines and is adding varicella in January 2026. Germany recommends 14 and added meningococcal B in January 2025. Canada recommends 16. Australia recommends 15. 

Even Guinea-Bissau, a low-income West African country, recommends 12 vaccines for all children. Denmark’s schedule is the most minimalist of any developed nation. Calling it “international consensus” requires ignoring most of the international community.

The Høeg-Kulldorff assessment acknowledges that Denmark maintains high vaccination rates through trust and education rather than mandates. What the authors seem not to recognize is that systematically casting doubt on vaccines, by moving them to categories that signal uncertainty where none exists, is precisely how you erode that trust.

The case-by-case unraveling

Consider the vaccines now relegated to shared decision-making.

Rotavirus caused 55,000 to 70,000 hospitalizations and 20 to 60 deaths annually in the United States before vaccination. 

The Høeg-Kulldorff assessment dismisses this, noting only 3.3 deaths per year with rotavirus on the death certificate between 1999 and 2005, and 1.6 deaths per year after vaccines were introduced. They attribute this reduction to possibly “improved medical care, changes in diagnostic practices, or random fluctuations.” They do not mention the 85% to 95% reduction in hospitalizations. 

They cite the rare risk of intussusception (1 to 2 instances of a telescoping of the intestines per 100,000 vaccinated children) without contextualizing it against the tens of thousands of hospitalizations prevented, and without noting that wild-type rotavirus infection itself may trigger intussusception, meaning the vaccine might actually prevent naturally occurring cases. 

Before the vaccine, rotavirus was responsible for approximately $1 billion in annual direct and indirect costs. By moving this to shared decision-making, the administration is transferring that cost back to families and the health care system under the guise of “choice.”

Influenza killed 288 American children in the 2024-25 season, the deadliest in 15 years. Among those with known vaccination status, 89% were not fully vaccinated. 

The Høeg-Kulldorff assessment cites a 2018 Cochrane review finding that flu vaccines reduce influenza infections in children but that trials could not evaluate mortality because deaths were too rare in study populations. This argument exploits a statistical impossibility: because death from influenza is a rare event in healthy children, a randomized trial would require millions of participants to demonstrate a mortality benefit, a scale that has never existed for any pharmaceutical product.

Hepatitis B has been universally recommended at birth since 1991. Since then, childhood hepatitis B infections declined by 99%. The birth dose catches infants whose mothers were not screened, whose screening results were not transferred, or whose infection status changed after testing. When infants are infected at or around birth, 90% develop chronic infection, and 25% of those will die from cirrhosis or liver cancer. (See my previous CIDRAP op-ed on the subject.)

Meningococcal disease is rare but among the most devastating infections in medicine, with a 10% to 15% case-fatality rate, death within 24 hours of first symptoms, and survivors left with amputations and brain damage. The assessment notes that US incidence is 0.12 per 100,000 and cites a World Health Organization (WHO) threshold suggesting mass vaccination for rates above 2 per 100,000. This citation is epidemiologically misleading: the WHO threshold was designed for reactive mass campaigns during epidemics in the African Meningitis Belt, not for routine preventative care in high-income nations. 

The United Kingdom, Ireland, Italy, the Netherlands, Portugal, Australia, Canada, and Germany all have routine meningococcal programs despite low incidence. Germany actually expanded its program in January 2025 to include meningococcal B. Denmark is the outlier, not the consensus.

A clear pattern

The trajectory was visible months before the January decision. In May 2025, Secretary Kennedy announced that COVID-19 vaccination would be removed from the routine schedule for healthy children and pregnant women, a decision made without ACIP input. In September, ACIP formalized shared decision-making for COVID-19 across all age-groups and removed the routine preference for the measles, mumps, rubella, and varicella (MMRV) combination vaccine. In December, the committee voted to end the universal recommendation of the hepatitis B birth dose. 

It creates the appearance of patient autonomy while erecting structural barriers to care.

Yesterday’s decision consolidates and extends this approach across the schedule.

The playbook is now clear: invoke shared decision-making not as a scientifically grounded response to genuine uncertainty but as a mechanism for stepping back from vaccines without formally withdrawing them. It creates the appearance of patient autonomy while erecting structural barriers to care. It uses the language of nuance to accomplish blunt ideological goals.

The states’ response

The divergence between federal guidance and scientific consensus has prompted unprecedented action. The West Coast Health Alliance, comprising California, Oregon, Washington, and Hawaii, was formed explicitly to provide evidence-based recommendations independent of CDC. The Northeast Public Health Collaborative encompasses eight states pursuing similar goals. California has passed legislation authorizing its public health department to base recommendations on independent medical organizations’ advice rather than on ACIP’s.

This fragmentation is itself a harm. Vaccine policy works best when it is consistent, clear, and backed by institutions the public trusts. When states must route around federal guidance, when different regions of the country operate under different recommendations, when the message to parents varies by zip code, confusion multiplies and confidence erodes.

The administration frames its changes as restoring trust. The WHO’s director of immunization, Kate O’Brien, MD, MPH, offered a different assessment: “I don’t know of another example of where a country has dramatically, substantially changed their immunization program.”

A call for clarity to protect children

The American Academy of Pediatrics has announced it will continue to recommend the full complement of childhood vaccines and called the changes “dangerous and unnecessary.” The American Medical Association warned, “When longstanding recommendations are altered without a robust, evidence-based process, it undermines public trust and puts children at unnecessary risk of preventable disease.” 

Professional societies, including the American College of Obstetricians and Gynecologists, the American Academy of Family Physicians, and the Infectious Diseases Society of America, continue to support these vaccines.

Clinicians navigating this landscape should be clear-eyed about what has happened. Recommendation categories have been manipulated for purposes beyond their design. The evidence for rotavirus, influenza, hepatitis B, and meningococcal vaccines has not changed. What has changed is the willingness of federal officials to let that evidence guide policy.

Shared clinical decision-making was designed to acknowledge complexity where it exists. Using it to manufacture complexity where none exists is a betrayal of the concept and, ultimately, of the patients it was meant to serve.

Dr. Scott is a clinical associate professor of infectious diseases at Stanford University School of Medicine and first author of “Updated Evidence for Covid-19, RSV, and Influenza Vaccines for 2025–2026” in the New England Journal of Medicine.